The microbes that live in the human gut microbiota possess novel and ill-defined metabolic capabilities. Many of the metabolites produced by this microbial ecosystem are absorbed by the human host and ultimately excreted by the kidneys. Such solutes accumulate when the kidneys fail and comprise a significant portion of the uremic solutes found at very high level in the plasma of patients maintained on dialysis. p-cresol sulfate (PCS), indoxyl sulfate (IS), and trimethylamine-N-oxide (TMAO) are three prominent uremic solutes that depend upon microbial metabolism of diet-derived molecules. Each exhibits inter-individual variability in the quantities produced, suggesting that microbiotas differ in their ability to produce these molecules. PCS and IS have been associated with poor outcomes in renal patients and TMAO has been linked to cardiovascular events in humans. The ultimate goal of this research is to understand how the human gut microbiota may be modulated to decrease the production of these compounds. The main objectives are to elucidate (i) the microbial genes and species that are the key contributors to PCS, IS, and TMAO production; (ii) the impact of diet on the production of these solutes; and (iii) the best method for reprogramming a high-producing microbiota to a low-producing phenotype via administration of other gut-derived species. In Aim 1 a new machine learning software, ClusterFinder, will be used to query ~900 sequenced gut microbiota genomes to predict genes and gene cassettes that contribute to PCS, IS, or TMAO generation. Predictions will be validated in pure culture and in vivo, using a gnotobiotic mouse model in single and multiple species colonizations. Gene predictions will be genetically validated using gene deletion or heterologous expression. In Aim 2 healthy omnivorous humans with stable high and low urinary TMAO, PCS and IS production will be identified. Microbiome-encoded genes and taxa associated with solute production phenotypes will be determined by analyzing stool samples for these individuals using 16S rRNA-based microbiota enumerations, metagenomics, and metatranscriptomics. Aim 3 will address whether diet affects the production of TMAO, PCS, or IS using either gnotobiotic mice colonized with bacterial consortia validated in Aim 1, and a humanized mouse model, consisting of ex-germ-free mice colonized with the human fecal microbiota identified in Aim 2 of high- and low-producers of TMAO, PCS, or IS. Diets will include high vs. low fiber, high vs. low protein, or L-carnitine supplemented vs. vegetarian. Decreases in TMAO, PCS and/or IS production will be associated with changes in microbiota composition and function. Aim 4 will identify the most effective method for microbiota reprogramming in humanized mice to decrease TMAO, PCS and IS production using transplants of an intact microbiota, donor microbiota generated culture collections, or type strain representatives. The use of antibiotics before transplant and the influence of dietary reinforcement will be tested as methods of aiding microbiota reprogramming.